>
Experimental Procedure
> Theory
> Results and Discussion
>
Conclusion
> Acknowledgment
> References
P.
Leisner, G. Bech-Neilsen & P. Moller,
Centre of Advanced Electroplating,
DTH 425, DK-2800 Lyngby, Denmark.Introduction
It is mostly
desirable to obtain a good throwing power in electrodeposition
processes. Particularly in through hole plating of printed
circuit boards (PCB's) a uniform distribution of the deposited
copper is demanded. The desired throwing power is obtained
in traditional production by use of additives. However,
these additives represent a considerable part of the costs
in the PCB plating process. Therefore, it is beneficial
from an economical, but also from a process engineering
point of view, if the desired throwing power can be obtained
by applying pulse plating instead of using additives. Safe
handling of waste from production will also benefit from
the absence of additives.
Puippe and Ibl (1)
have shown that the throwing power can be improved by applying
very high frequency square waves (> 1 MHz). This process
utilizes the capacitance effect, that is, the discharge
of the cathode in the off-period runs the deposition process
so that no Ohmic drop in the electrolyte will decrease the
throwing power. White and Galasco (2)
have, under pulse reversal current (PR) conditions in an
additive-free bath, obtained surface distribution comparable
to those obtained by DC plating with additives. The frequency
was between 17 and 50 Hz.
The aim of this project was to improve
the throwing power by combining periods of electrodeposition
and periods of electrodissolution in a low frequency (1<Hz)
PR process. The application of low frequency is important,
if the process should be used on an industrial scale.
Experimental Procedure
Copper was deposited from a commercial
acidic plating bath without additives (Cu20 g/l; H2SO4200
g/l). The bath was operated at room temperature under air
agitation. Phosphorus depolarized copper was used for anodes.
Commercial equipment was used to supply pulsed current (AXA
30 V/100A).
In the first part of this project
copper was deposited on Assaf test panels (2),
which are very useful for investigating the throwing power.
The test panel consists of a quadratic (4 x 4 cm2) stainless
steel cathode mounted at a distance of 5 mm from a plastic
board. The panel was placed facing the anode. Following
the plating the thickness was measured at the center of
both sides of the cathode with an X-ray equipment (Fischerscope
X-ray 1550). The current efficiency was determined by weighing.
In the second part of this project
through hole plating was carried out on PCB's. The hole
to land thickness ratio was measured for holes with an aspect
ratio of 2.0 (0.8 mm holes in 1.6 mm boards) using microscopy.
The results are presented as the average of 14 holes.
Theory
In electroplating from acid baths,
where the current efficiency is close to 100%, the material
distribution will be proportional to the current distribution.
Thus, the throwing power can be expressed by Wagner's number,
Wa, which describes the current distribution during a plating
process:
where k is the electrolyte conductivity (S/cm), L is the
characteristic length of the process (cm), and dn/di is
the slope of the polarization curve. For small values of
Wa (Wa <<1) the current distribution is dominated
by the Ohmic drop in the electrolyte and, therefore, uneven,
Conversely, the current distribution is dominated by the
activation resistance at high values of Wa (Wa >>1)
and therefore more even.
A more uneven current distribution
in the anodic period compared with the cathodic period is
needed in order to obtain a leveling effect during PR plating,
Hence, Waa < Wac, which can be formulated as
In the Tafel-region this can be written as
where b is the Tafel-slope and subscripts a and c refer
to anodic and cathodic conditions, respectively.
Of course, the ratio between the amount
of copper dissolved in the anodic period and the amount
of copper deposited in the cathodic period is also an important
parameter, which can be stated as:
Results and Discussion
The polarization curve for a RDE in
the plating solution is shown in figure 1. From this curve
the Tafel-slopes are found to be bc = 95 mV and ba = 53
mV, which is in fair accordance with the model suggested
by Mattsson and Bockris (4)
and further developed by Reeve
(5). Inserting these values in eq. (3),
one obtains
Consequently, an improved throwing power will be obtained
when eq. (5) is
fulfilled and both the cathodic and the anodic reactions
are in the Tafel-region.
The validity of eq. (5)
is supported by the results in table 1, which show that
PR plating applying the same current distribution in the
cathodic and the anodic periods result in the same back
to front thickness ratio as DC plating.
The results from plating on Assaf
panels is shown in figures 2-5. The throwing power as a
function of Qa/Qc, ia, Ta and ic has been investigated.
The results show that the throwing power is improved with
increasing Qa/Qc and to lesser degree with increasing ia.
At very high Qa/Qc values a "reverse" throwing
power can be obtained (the deposition is thickest in the
low current density area). The dependance of Ta is minor
as long as passivation is avoided. The throwing power passes
a minimum, when ic is increased. This can be explained by
a transition from secondary to tertiary current distribution.
At secondary current distribution condition Wa decreases
with increasing current density, but increases with transition
to tertiary current distribution. It is seen that the transition
in current distribution is influenced by Qa/Qa and that
copper passivates when too high anodic current density is
applied for too long time. The last phenomenon is illustrated
on Figure 6, where the upper limit for avoiding passivation
is shown as a function of ia and Ta.
The current efficiency of copper electrodeposition
on the Assaf panels is between 95 and 100% under both PR
and DC conditions.
The results from plating on PCB's
is in accordance with the results from the Assaf panels
(figs. 7-9). Again, the relative thickness of the deposit
in the low current density area is increased with Qa/Qc
and to a lesser degree with ia, and the dependence on ic
is influenced by Qs/Qc and ia.
Under DC conditions (i = 1,8 A/dm2)
the hole to land thickness ratio is found to 93%, which
is more than expected. This may be related to a relatively
low aspect ration.
Conclusion
It has been shown that the throwing
power can be controlled in an acid copper bath without additives
by combining periods of electrodeposition with periods of
electrodissolution in a low frequency PR process.
The most important parameter for increasing
the relative thickness in low current density areas is Qa/Qc,
but ia is important too. The copper passivates at high ia
and further copper dissolution occurs under formation of
pitting. The upper limit for avoiding passivation depends
on Ta. The relationship between the throwing power and ic
is complex, while the characteristics of the current distribution
is changed from secondary to tertiary distribution with
increasing ic. This means, that the throwing power passes
through a minimum with increasing ic. The change in distribution
characteristics is influenced by Qa/Qc and ia. The current
efficiency is close to 100% in all experiments.
The mechanical properties of copper
electrodeposition under PR conditions have not been investigated
in this project.
Acknowledgment
We are grateful to Mr. Anthony McNelly
and Mrs Anette Christiansen, Poxy Print, Nysted (DK), for
their collaboration and interest in this project.
References
(1) J.Cl. Puippe and N. Ibl, J. Appl.
Electrochem. 10 (1980) 775.
(2) J.R. White and R.T. Galasco, Plat.
Surf. Fin 5 (1988) 122.
(3) Y. Assaf, Plat. Surf. Fin. 10
(1980) 12.
(4) E. Mattson and J.O'M. Bockris,
Trans. Far. Soc. 55 (1959) 1586.
(5) J.C. Reeve, An Investigation of
the Pseudo-Steay-State Kinetics of Copper/Cupric-ion Electrode
in Dilute Aqueus Solutions of Sulphuric and Perchloric Acids,
Thesis, University of London 1971.
